search

jbae11 / depletion_rom

BSD 3-Clause "New" or "Revised" License
0 stars 4 forks source link

prof. Huff review: #16

Closed jbae11 closed 5 years ago

jbae11 commented 5 years ago

Please review the paper and let me know what you think!

katyhuff commented 5 years ago

I started this weekend - I need a bit more time. I spent 20 minutes just getting it to build... see new issue on that topic #21, as well as corresponding PR #22 .

jbae11 commented 5 years ago

thank you, I merged the PR and checked build on the master.

jbae11 commented 5 years ago

@katyhuff I asked the folks at ORNL and they said that there's nothing I need to do on their side, as long as the paper had no influence from ORNL. I added the acknowledgement and my `current address' block. Pending your review and approval, I plan to submit it after completing #20. Let me know what you think. thanks!

katyhuff commented 5 years ago

@jbae11

Re:ORNL, I think we discussed how it may be beneficial to your career to list your future place of employment (then, it can be included in your productivity metrics). Obviously, ORNL didn't contribute to the work so far, but perhaps after the review comments come back, you'll spend some time working on it while you are employed at ORNL. If so, then you may want to reconsider your stated affiliation.

Re: review. I am finding the review difficult because the paper has a lot of grammatical and clarity issues. Additionally, I do still feel that the work doesn't show much without a reserved test set for demonstration of accuracy. I will send many more comments on this when I get a chance, but I hope you will be patient. They key mistake is that you state that you validated the trained ANN with data that was used to train the model. There is no validation in such an activity. In fact, it's quite distressing that the results aren't exactly perfect (where are the deviations coming from? Numerics?). I haven't read far enough to tell because there are so many grammatical and clarity challenges ( subject verb mismatches, sentences missing words, "etc" without a period, misuse of words like "proprietary"), and clarity issues (failure to cite the various software mentioned, lots of passive voice)... etc. That is all to say that I need at least another week with this paper.

katyhuff commented 5 years ago

I will submit a PR fixing lots of these things -- please do not send me a new version, as I'm halfway through this one.

jbae11 commented 5 years ago

@katyhuff ping

jbae11 commented 5 years ago

@katyhuff Do you have any updates on this? Please let me know. thanks

katyhuff commented 5 years ago

@jbae11 My scribbles are here... I've not had a chance to make the edits. If you'd like to start, there is plenty to do. But, I still think the work is fundamentally flawed, since no demonstration set was retained. Additionally, I think the method should be discussed, mathematically. This feature is missing from the work. Perhaps we should have a call.

2019-bae-depletion-rom.pdf

jbae11 commented 5 years ago

@katyhuff Okay I am running the case with a test set reserved (definitely not on ORNL time, of course). Also, I just wanted to clarify that I'm making the edits, so there's no duplicate work? If that's the case, I'll make the PR soon. I think it'd be best if we have a call after the first round of edits?

thanks!

katyhuff commented 5 years ago

It's up to you. Certainly, if you're willing, it'll likely be much faster if you do it, rather than waiting on me to get to it. (Thank you!)

On Wed, Mar 27, 2019 at 8:17 PM Jin Whan Bae notifications@github.com wrote:

@katyhuff https://github.com/katyhuff Okay I am running the case with a test set reserved (definitely not on ORNL time, of course). Also, I just wanted to clarify that I'm making the edits, so there's no duplicate work? If that's the case, I'll make the PR soon. I think it'd be best if we have a call after the first round of edits?

thanks!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/jbae11/depletion_rom/issues/16#issuecomment-477405929, or mute the thread https://github.com/notifications/unsubscribe-auth/AAYCq1oxDMMIIqAXM6wIlIH3BAucXUfFks5vbBgXgaJpZM4bBiCi .

-- http://katyhuff.github.com

jbae11 commented 5 years ago

okay, I'll make time this weekend and make a PR! Thanks

jbae11 commented 5 years ago

@katyhuff I accidentally pushed all the changes to the master branch, but I made a huff_review.tex for all the questions written in the review. All the non-mentioned comments are accepted and reflected in the paper. Please let me know what you think. Thanks!

Also, Kausnik Banerjee wanted to see the paper, is it okay if I send him this draft?

jbae11 commented 5 years ago

@katyhuff Have you had the time to look at this? thanks

katyhuff commented 5 years ago

I have not. I can aim to take a look sometime this week. A response will not be possible before April 20th, though. (you're welcome to show Kaushik, of course... and again, I would recommend that if you're spending time at ORNL on the work, you should re-list your institution and include appropriate acknowledgment in addition to what already appears.)

jbae11 commented 5 years ago

I talked to my mentors in ORNL about the acknowledgement and they said that the current way was okay with them, as long as I dont spend work time on this work. Please let me know when you get the time to look over the paper. thanks.

katyhuff commented 5 years ago

@jbae11 Thank you for your patience*. My edits are attached. I'll close this issue (because the review is complete) and open a new issue summarizing the remaining concerns. That issue can be closed with a PR that addresses the comments in the pdf.

bae-round-2.pdf